How to deduplicate a string where it only deletes consecutive duplicates - c++

Question

The goal of the program is to take a string like "kcck" and delete the consecutive duplicates. It should first iterate through the string and delete cc leaving kk; then go through again and delete kk; then return "empty" since there are no characters left in the string.
Another example, "aabggtcc" should return "bt".
int i;
int j = i+1;
string deduplicate(string input) {
for(i=0; i<input.length(); ++i) {
while(j <input.length()) {
if(input[i] == input[j]) {
input.erase(i);
input.erase(j);
}
else if (input[i] != input[j]) {
++i; ++j;
}
if(input[i] == '\0') {
cout<<"empty";
}
}
}
return 0;
}
int main () {
cout<<deduplicate("aabg")<<endl;
cout<<deduplicate("ag")<<endl;
cout<<deduplicate("btaabb")<<endl;
return 0;
}
When I run the code it gives me:
libc++abi.dylib: terminating with uncaught exception of type std::out_of_range: basic_string

Answer 1

There are couple of issues with your snippet,
deduplicate function is returning zero(0) all the time
j is initialized in global scope and is never reset for new string
As you erase std::string::length() is calculated on new string hence your index i and j are won't point to same laocation.
Here is the snippet with rectified error,
string deduplicate(string input) {
int i = 0;
int j = 0;
while (i < input.length()) {
j = i + 1;
bool isRepeated = false;
while (j < input.length()) {
if (input[i] == input[j]) {
input.erase(j,1);
--j; //as string length is reduced by 1
isRepeated = true;
}
++j;
}
if (isRepeated) {
input.erase(i,1); //remove first letter as well
--i;//sting length is reduced by one
}
++i;
}
return input;
}
int main() {
std::cout << deduplicate("aabg") << endl;
std::cout << deduplicate("ag") << endl;
std::cout << deduplicate("btaabb") << endl;
return 0;
}
output:
bg
ag
t
which can be even simplified as,
std::string deduplicate(std::string input) {
std::string s ="";
for (auto c : input) //loop through all char
{
int f = 0;
for (auto c1 : input)
{
if (c1 == c)
{
f++; //increment if char is found
}
}
if (f == 1)//append char only if it present ones
s += c;
}
return s;
}

Answer 2

You are decreasing the size of the string whenever you call string.erase() that's why the variable i eventually exceeds the "current" string size input.length() in the while loop, and you get an error std::out_of_range: basic_string when you try to access input[i] in the if and else if conditions of the loop.
Try to manually go through the loop on the string on which you got the error and you will see that i has gone out of bound (i.e. i >= input.length()) in the while loop

Answer 3

With C++11 and on, instead of iterating over each character and making the comparison manually, you can use std::basic_string::find_first_not_of to look forward from a position in the string and find the first character not of the current character. If the position returned by .find_first_not_of is more than 1 from the current position, you can use .erase to erase that number characters. If the return is 1, then just increment your current position and repeat.
To operate on all duplicates characters in the modified string, you simply wrap it all in an outer-loop, keep a copy of the string before before entering the inner-loop to remove duplicate characters, and compare if the modified string is equal to your copy or the .length() is zero for your exit condition.
You can do something similar to the following:
#include <iostream>
#include <string>
int main (void) {
std::string s;
while (getline (std::cin, s)) {
std::string current;
do {
size_t pos = 0;
current = s;
while (pos < s.length()) {
size_t duplicates = s.find_first_not_of (s.at(pos), pos);
if (duplicates != std::string::npos && duplicates > pos + 1)
s.erase(s.begin() + pos, s.begin() + duplicates);
else if (duplicates == std::string::npos &&
(s.end() - s.begin() - pos) > 1)
s.erase(s.begin() + pos, s.end());
else
pos++;
}
} while (current != s && s.length());
std::cout << "'" << s << "'\n";
}
}
Example Use/Output
$ echo "kcck" | ./bin/ddcpp
''
$ echo "aabggtcc" | ./bin/ddcpp
'bt'
$ echo "aabg" | ./bin/ddcpp
'bg'
$ echo "ag" | ./bin/ddcpp
'ag'
$ echo "btaabb" | ./bin/ddcpp
'bt'
There are a number of ways to approach the problem and as long as they are reasonably efficient there isn't any one right/wrong way. If you have a modern C++ compiler, letting some of the built-in container functions handle the work is generally a bit more robust than reinventing it on your own. Look things over and let me know if you have questions.